PY-GC/AED and chemometric correlation to characterize sewage sludges of different origins

The solid organic fractions of sewage sludges originating from food-processing and domestic (rural and urban) depuration systems of Lorraine (Northeast of France) were characterized by flash pyrolysis coupled with gas chromatography/atomic emission detector (Py-GC/AED). These analyses generated a high amount of data which could not be processed without a further step using a chemometric approach. Thus, a heuristic methodology was developed, relying on the application of pattern recognition methods, such as Principal Components Analysis, hierarchical classification and the k-means method generally used in chemometry. In order to avoid subjective selection and reduction of data, a signal treatment relying on the alignment and the segmentation of pyrograms was performed before to the statistical analysis. The gathering of test-samples analysed several times validated this methodology. The main chemometric results provided a typology of sludges which consisted in a classification in different groups, and allowed to display the characteristic peaks of each group. This chemometric analysis gathered a group of sludges with an average chemical composition characterizing stable organic material and pointed out the atypical chemical composition of several sewage sludges. These results demonstrated the efficiency of Py-GC/AED associated to chemometric methods for the global characterization of the insoluble organic fraction of sewage sludges. It could be possible to apply, with further calibration, to the monitoring of sludge composition during its treatment. (C) 2003 Elsevier B.V. All rights reserved.